I stopped by the living computer museum a couple weekends ago and found some amazing old unixy systems that are older than I am. I want to try to run some non-trivial program on them so I figured I'd code it up first.
I'm going to try a program that solves the towers of hanoi problem in a primitive version of C.
While writing this I figured it would be helpful to actually emulate some old machines to see how old C actually runs on old hardware. I investigated some hardware emulators
Configuring and developing the software was much different than it is today.
After using the command line without command history I started to make my filenames very short. Now I know where cp, mv, du, etc. came from.
Having to scope all variables at the top of the function means you may have to reuse the variables for different purposes. You might as well call them i, j, k. It makes sense that early versions of K+R C manuals consistently used these letters as variable names.
I am so used to compiler help: typing, mismatched arguments, error messages that mean something, correct line numbers, etc.
More than once while editing hit q expecting changes to be saved before exiting. They were not. There was no warning.
Source control: I can't believe progress was once made without this tool.
All of these issues make me very gun shy about any changes to the system. I started composing all of my actions before actually entering any of them. I felt like I was chisleling a marble statue. Mistakes are costly. This was a bad thing but it wasn't a 100% bad thing. I think keeping more of this think before doing attitude in modern environments would be a good thing.